Wednesday, October 30, 2013

2013 Retrodiction: How Player Metrics Predicted Wins

As the 2014 season starts, let's not forget the past and what we can learn. NBA player metrics are more popular than ever, but which ones do we trust? How useful are they? One method for evaluating metrics is to use the past values for each player to predict team wins in the current season. For example, for the Lakers we use Dwight Howard's PER in 2012 along with his minutes in 2013, we use Kobe Bryant's PER in 2012 along with his minutes in 2013, and so on until we have every player.

The metrics
PER: John Hollinger's invention. Explained here. Uses box score stats to create a player per-minute productivity value. Values usage highly. Adjusted for pace.
Win Shares: Uses box score stats to calculate individual wins. A large team defense factor is evenly divided among players. Values efficiency.
Wins Produced: Like Win Shares, uses box score stats to calculate individual wins. Instead of a large team defense factor, heavily values rebounding. Also values efficiency.
RAPM: Ridge-regression adjusted plus minus. Fundamentally based on whether or not a player's team scores more often or allows less points when he's on the court. Adjusts for pace, strength of schedule, among other factors. The "ridge" part regresses heavily to a prior value based on possessions, i.e. it's hard for a player to move away from his prior if he rarely plays.
-npi RAPM: Non-prior informed RAPM. The priors are all set to 0. 
-Vanilla: Uses the previous three seasons of npi RAPM. Most recent seasons are weighed more heavily.
-RAPM: The prior is the previous season's RAPM. The starting point is the first dataset used (2001.)
-xRAPM: A mix of RAPM and a statistical plus/model using box score stats and height.

To rate the various metrics, we can use what's known as a root-mean square error. Sum all the squared differences between wins and predicted wins, and then take the square root. The effect of this is to penalize the biggest errors, and then to calculate the root mean (the final units are in wins, not squared wins.)

Which metric won the 2013 retrodiction? xRAPM blew away every other metric. A root-mean squared error under 6 is extremely good for a prediction for the regular season. Although knowing the minutes distribution is a huge advantage, it's worth noting that the best analysts and Vegas topped out at 5.94 for the 2013 season. (Interestingly, while PER did not perform well, Hollinger's own predictions did.) For the worst performing metric, that is Wins Produced. Only PER approaches how poorly Wins Produced did, but that's not with the "real" wins. Most metrics had minor differences in how they predicted real wins and Pythagorean wins, but PER "lucked out" that real wins trended much closer to its own prediction. Win Shares does extremely well and beats out the other adjusted plus/minus methods, but we won't stop here.

 ........
PER
Win Shares
Wins Produced
npi RAPM
Vanilla RAPM
RAPM
xRAPM
Real wins
7.00
6.62
7.95
7.51
7.27
6.74
5.67
Pyth. wins
7.98
6.93
7.75
7.41
7.23
7.11
5.85


Since you're really not testing anything on teams with low roster turnover, I used the percentage of minutes from new players excluding rookies to calculate squared errors at three points: no new players, half the minutes from new players, and a completely new team. Why is this important? Box score metrics assume defense is explained by rebounds/blocks/steals, and since a good defender will force more missed shots (i.e. rebounds), it's hard to test this without the player switching teams.

The results are shown in the table below. The most relevant line is 50% because the team with the most roster turnover topped out at 59.3%. The "Win" metrics do very well when the roster doesn't change, which isn't surprising because their metrics are built from offensive/defensive team measures, but they start to perform horribly with heavy roster change. Weirdly, PER doesn't change much when the roster does for real wins; there was something kooky going on with PER trying to explain teams with roster change (case in point: the Rockets gained Asik, who PER underrates because he's a defensive player who doesn't shoot well, but the Rockets underperformed their point differential.)

The RAPM metrics are rough in estimating wins on teams with no roster turnover, but this makes sense: plus/minus models are known for their noise. However, they do much better with roster turnover, especially the metrics that use priors: xRAPM and RAPM. This is despite completely whiffing on the Lakers, which the box score metrics did not do, but that team struggled with injuries and chemistry.

PER does the best at 100% minutes from new players, but, again, the 50% line is more relevant and it's more the case that PER was horrible at teams with no roster change, making them look better when projected out to 100. As a last note, non-prior forms of RAPM are generally disregarded, but they still appear to be as predictive as widely used box score metrics like Win Shares.


% of mins. from new players
PER
Win Shares
Wins Produced
npi RAPM
Vanilla RAPM
RAPM
xRAPM
Real wins
0
41.3
-3.6
3.1
19.7
16.3
24.8
16.4
50
53.5
71.4
98.3
77.7
74.1
57.4
41.4
100
65.6
146.3
193.4
135.7
131.9
90.0
66.4
Pyth. wins
0
43.7
0.2
15.7
21.7
26.2
22.1
20.6
50
75.3
75.9
85.9
74.1
67.4
67.0
42.2
100
106.8
151.7
156.2
126.5
108.7
111.9
63.8

How to translate the metrics to wins

For Win Shares and Wins Produced, the work has already been done. Just multiply by minutes.

PER has a companion stat for wins: EWA (estimated wins added.) It's explained here:
VA: Value Added - the estimated number of points a player adds to a team’s season total above what a 'replacement player' (for instance, the 12th man on the roster) would produce. Value Added = ([Minutes * (PER - PRL)] / 67). PRL (Position Replacement Level) = 11.5 for power forwards, 11.0 for point guards, 10.6 for centers, 10.5 for shooting guards and small forwards

EWA: Estimated Wins Added - Value Added divided by 30, giving the estimated number of wins a player adds to a team’s season total above what a 'replacement player' would produce.

Plus/minus stats are translated to wins using the Pythagorean method. The formula is points scored^14/(points scored^14+points allowed^14). The expected plus/minus of the team is translated into points scored versus points allowed relative to the league average.

Rookies are assumed to be heavy negatives. The average value for rookies was estimated from past values.
WS/48 mins: 0.05
PER: 13
WP/48 mins: 0.05
Plus/minus (RAPM's): -1.96

Tuesday, October 29, 2013

2013-2014 NBA Season Preview: Eastern Conference

It's that time again: a new season dawns, and with the first games looming I'm publishing the first part of my 2013-14 preview. While there's plenty of content about basketball online already looking at the upcoming season, I've added a lot of unique stats and graphics you won't find anywhere else. It's a (somewhat) brief guide to every team and possible evidence of my incompetence when the predictions go awry.

Now in reverse order from best projected record to worst....


Philadelphia 76ers

I think Philly is being underrated for two reasons: Thaddeus Young is underrated, and offensively they don't have a number one option, which is a component people overemphasize in team strength. Alas, when people predict you'll win 15 games, it's easy to be underrated -- there's no question this is a bad team; I just think they'll win more than 15. Last year they were nearly league average on defense. But some of that is thanks to Jrue Holiday, who won't be with the team anymore, and a lot of that is due to effort -- which shouldn't be close to 100 percent this season.

While filling out win projections, one normally has to do minutes projections (this is one of the under-reported segments of NBA analytics, as a minutes projection can destroy your prediction, and it's extremely valuable to know how many games a player can undertake in a variety of situations like post-injury.) For most teams, this was fairly easy, and I didn't incur problems soaking up the bench minutes. However, the 76ers were another story. Simply put, who are these guys, and are we sure this isn't a D-league team with Turner/Young/Hawes thrown on top? Vander Blue and Hollis Thompson sound like computer-generated names you'd see on NBA2K. Filling out the rest of the team's minutes is arduous in this case because it's likely you're going to get lot of minutes from "below replacement-level" guys, if you will, but some of these players probably even aren't on the roster yet.

To make matters worse, the 76ers have decided not to play Noel for the entire season. The reactionary comments condemn the team for trading away an all-star point guard for an injury-plagued big man, but that's not accurate and misses some subtle points. Also, the latest news is that there was no change in his recovery; they really have just decided to tank. The risk of injury in the NBA draft can be overvalued. The fundamental question comes down to risk versus reward, how much a player can help even if the minutes are limited, but it's more simply this: would you rather have Brandon Roy or Randy Foye? Foye can give you a large quantity of NBA minutes, and while he's not a star he's not a bust either; he's been a rotation player for years. Yet Roy gives you a level of talent and production far greater than Foye. Obviously Roy is the better choice, even though he retired early, although the question is complicated from cumbersome, large NBA contracts (perhaps with Bynum largely non-guaranteed deals will become more common with injury risks.) While critics point to Oden's odyssey, Blake Griffin sat out his "rookie" season to emerge the next year as a fully-formed 20/10 power forward. The defensive potential of Noel, whose stats were arguably better than Anthony Davis on defense, is enough to offset his draft slot (seventh) in a weak draft and a swing-for-the-fences rebuilding move. Noel's stats and measurements suggested a player with a high probability at being a great player, and had a clear degree of separation from the field. Given the injury concern, he wouldn't be worth taking over a player of the same caliber, but to completely write him off in favor of lukewarm prospects is ridiculous.

As for a discussion of guys who will actually play, there's a brief list of notable players. This is Evan Turner's year to shine because it's as wide-open an opportunity as anyone will get. They will encourage him to shoot, even if he's not entirely accurate, because they have no other options. He's taken on the tired offensive strategy of long two-point bombing without the accuracy to make it viable. Right now his offense is putrid, but if he can either increase his range or change his attack focus, going to the rim where for two seasons he shot near 65%, he can push his shooting percentages into an acceptable range. But he doesn't offer anything else of value, with the odd exception of his rebounding. His rebound percentage was 10.2%, nearly 20% on defensive rebounds, which would be good for a power forward, much less a perimeter player who's often a shooting guard. And the previous season he was at 12.3% -- that's LeBron James territory. The list of guards who have accomplished this is small, and it's a quirky group of guys, some you wouldn't expect -- from Magic to Fat Lever, from Kidd to Dorell Wright. Unfortunately, a green light to shoot may submarine any effort to improve his efficiency, and the once highly coveted prospect is seeing his future disappear from under his feet. He's also a likely trade candidate.

On a brighter note, Thaddeus Young will continue his run as an underappreciated piece. He's definitely their best player now, unless someone breaks out, and no, that's not a great sign. He's a two-way force who doesn't leap off the page in one single category, but he's a fantastic finisher and is hard to keep out of the paint. Defensively, he should help hold down the fort on a disappointing season, and given his young age, just 25, he's going right into his prime. A forward who plays both positions, his tweener status hasn't led to any trouble, in contrast to other tweener forwards, and in fact his positional flexibility is an asset. He deserves to play for a better team, either starting or going for the sixth man of the year award, but this is his purgatory, for better or worse. And the center holding down the middle, Spencer Hawes, is really the only other player of note. He's a stretch center who barely has three-point range who managed to make 36% on 87 attempts, and he's a surprisingly good passer who picks up a lot of assists. His rebounding is only okay, and beyond his shotblocking his defense leaves a lot to desire, but to a team as devoid of talent as Philly he's suddenly a vital piece. In fact, his outside shooting is sorely needed because the team is bereft of shooters, especially with Wright gone: he made 26% of their total three-pointers, and their top three shooters from outside by makes (and five of their top six) are all gone. It's going to be an ugly offense, and they'll need the NBA-level competence of Young and Hawes to keep things from reaching Washington Generals-level.

The 2014 season for Philadelphia may as well be computer generated, and with the stats-heavy front office it might even be a proposal. The only reason this season should exist is to give experience to their young guys and await their draft pick in the most intriguing draft since Oden/Durant. Jrue Holiday was thrust into the position of lead scorer and creator when that is clearly not his strength -- and now they don't even have him. The offense will be brutal; the defense, surprisingly, doesn't project as poorly, but it's difficult to get a read on a team that quickly goes from playoff hopeful to tanker -- just how much worse will they get? How do you judge that? They were not a terrible defensive team last year, but they've lost their head coach and their will to compete. But some of their pieces on defense are pretty good, like Young and Lavoy Allen, a player few fans could pick out of a lineup but nonetheless remains a pretty valuable on that end of the floor. The 76ers project as the league's worst offense -- in other words, they're actually going to miss Nick Young, and the shot clock could be one of their most intimidating competitors. But, hey, they could end up with Wiggins and Noel to start the 2015 season -- and that's all the fans will think about.

Random prediction: Evan Turner will be traded to the Bucks for either one of their big men and another small piece or for a first round draft pick (top five protected.)

Win/loss prediction: 20-62


Orlando Magic

Lost in the praise of unexpected gems Tobias Harris and Vucevic was Orlando's ineptitude. They had the second lowest win total and the second worst point differential. With Howard and Ryan Anderson gone, scoring duties were handed out to players unsuited for the role. It says a lot for Howard's value, and Anderson's, that the team imploded so horrifically when just before they were a perennial 50 win team with a trip to the finals. They had a bottom five offense, but it was particularly egregious with their top scorers. The graph below shows every player with at least 1000 minutes and a usage rate of 22.5. Glen Davis, amazingly, was the number one option, and it wasn't pretty. Tobias Harris, by the way, almost qualified: a 23.1 usage but only 973 minutes. A TS% of 52.4 with room to improve from behind the line gives them another option to lay off the pressure from role players like Davis and Afflalo. You may notice those marks aren't far from the league average, so how does that compare to everyone else?


Orlando has some of the worst top-line scoring talent in the league. Davis isn't even a heavy usage player, as it's common for teams to have players at 28 or more. Their lack of shot creation resulted in the worst FTA/FGA ratio in the league: 0.197. Orlando's players are all at the two worst edges of the graph: weak shot creators with a low usage and inefficient shooters. But are all poor offensive teams that way?


Looking at the ten worst teams in terms of TS%, where the Magic actually rated ninth from the bottom, even poor offensive teams have some decent first and second options. The lowly Bobcats have four players who rate higher in terms of TS% and usage. The worst offensive team, the 76ers, had a player who shot more than Davis with a higher percentage, although he (Jrue Holiday) is gone now. The only team with first options who are just as terrible was the Phoenix Suns.


The Magic did better by team TS% because some of their other options were fairly decent; the problem is that they didn't have a clear number one option. Redick and Ayon, for example, were fairly efficient, but they're gone; fortunately, the rookie Andrew Nicholson is back and he appears to have added legitimate three-point range in the off-season. Vucevic was slightly below average on TS% in Afflalo's range, but he can at least take pressure off Glen Davis on offense. He has a decent touch inside but needs to work on his footwork.

Speaking of Vucevic, the Dwight Howard trade was a category 5 hurricane of NBA destruction. Consider the devastation. Orlando loses Howard and receives gift-basket of assorted players including Afflalo, who doesn't find success in a larger role, and proceed to win 20 games. The Lakers got Howard, Earl Clark, and Chris Duhon. It was a disaster of a season, plagued by injuries and chemistry problems. When they rallied to make the playoffs, Kobe suffered a torn Achilles and missed the rest of the season, where the Spurs easily finished them off. Howard went into free agency and spurned the Lakers for a younger team in the same conference. While Denver seemed poised to emerge from the trade victorious, bringing in Iguodala to shore up their defense and only losing Afflalo, Harrington, and a conditional pick, they lost unceremoniously in the playoffs with homecourt advantage and Iguodala left for the Warriors. Philadelphia let their star go, the aforementioned Iguodala, but they received Bynum from the Lakers. Unfortunately, they endured a series of set-backs from the immature player and he never played a single minute; he left for Cleveland in the summer. The 76ers, to make matters worse, only received aging guard Jason Richardson in the trade, who's out with a major injury, and let Vucevic go to the Magic where he's been a pleasant surprise as a double-double machine. So yes, that's right, the only thing left standing in the rubble from this disaster is Vucevic with the Magic. No one saw that coming.

Sending out Redick, the Magic received the young Tobias Harris, who's been tormenting the Bucks after breaking out. He's a combo forward who slashes, but his rebounding is really impressive. Some may argue Harris and Vucevic put up empty numbers, but I think it's the opposite case: the team was 7th in DReb%, meaning if they were on an average team their rebounding numbers would increase. Given that there were no shot creators or first options, this makes their job harder on offense because they're left open less often and aren't being fed the best passes. Harris shot well on two-pointers but hit only 31.5% of his three-pointers despite taking over three per 36 minutes. Since he was only 20 years old, given natural improvements and a league average three-point percentage he can be a decent choice for a first option this season. Along this front the Magic have the rookie Oladipo. I don't believe he'll have a big impact right away, since it's hard for almost every new player in his first season and defense is more of his calling card, but given their depth they do need him.

The team is optimistic because Tobias Harris and Vucevic were pleasant surprises, but I would temper that enthusiasm. Orlando's defense was so terrible they missed Glen Davis (who due to his quick feet and girth was actually pretty good to his credit.) They won't to much better, and there's no reason to chase a better record. Of course, there's a chance a young player takes a major leap forward, like Harris, or that Oladipo turns into Wade, but in prediction we don't project the exciting, least likely things to happen. They'll be a little better but still terrible. But I think the fans are happy to escape the dark cloud surrounding Howard and coming through the trade disaster relatively unscathed.

Random prediction: With Burke injured and a thin cast in Orlando, Oladipo wins the Rookie Who Played the Most Minutes Rookie of the Year Award.

Win/loss prediction: 21-61


Charlotte Bobcats

The Bobcats, having been tired of toiling in absolute ineptitude for so long, used the strategy of aiming for mediocrity to dull the pain of losing 60 games, again and again, by splurging on "big" free agent Al Jefferson. While there's nothing necessarily wrong with that strategy given that the organization is tired of being a basement-dweller and want to float to respectability, the execution is what I'd criticize. First of all, the offense was horrible, but the defense was too, and now Al will take up a huge chunk of their center minutes, sending intriguing shotblocker Biyombo and aging defensive positive Haywood to the bench. Having a poor defensive center can be managed, especially if you're only aiming for 0.500 or less, but they drafted Cody Zeller high. If all goes well and they build around Cody Zeller and Jefferson remains healthy, that's their frontcourt for a long time -- Al is still only 28. Defensively, this frontcourt would simply be a lost cause, based on how short-armed Cody projects. That is not how a smart organization plans for the future. And it's not like they can do anything but base their team around Jefferson, since it's a poor team committing a lot of money for a big man who may not be easy to move in a trade. He's also the anti-Okafor, who spent his best years in obscurity as an underrated, valuable center, especially on defense, but he could never score in high doses. If the front office thinks Al Jefferson is a huge step up from Okafor, I'm afraid they will be mistaken.

Regardless, that's their future now, and they're entering yet another season where the playoffs look unreacable even in a shallow eastern conference, without a young marquee player to entice a fanbase. Michael-Kidd Gilchrist was called an extremely safe draft pick last year, with a high chance of being a very good player, but he was one of the worst shooters in the league, death for most perimeter players. In separating draft potential from player performance the Bobcats are a suggestion that the organization's development should be taken into account -- wouldn't he have done better on, say, the Spurs if given the minutes? The rest of the team is buoyed by Kemba Walker, who quietly had a fine season for a shoot-first point guard and Gerald Henderson, an athletic shooting guard who is somehow almost 26. For a random stat that highlights how overlooked Henderson is, his TS% and usage% (efficiency and volume) were almost exactly the same as Paul George's, off by only hundreds of a percentage point, and though he's not the defender George is he's still pretty good.

For a pure experiment, it will be interesting to see how Al Jefferson improves the team's offense since his skills -- shot creation at a low shooting percentage with shockingly few turnovers -- are attuned for a group of players who have trouble even putting up a shot, and it's not like the other guys on the team will make things easier for him. In a weird way, Al Jefferson's offense is underrated because the modern stat movement has focused on shooting efficiency, mainly TS%, while strangely ignoring turnovers, even though they are obviously damaging. The list of players who have a season of  >25 usage% (average is 20 and league leaders are usually above 30) and <6 TOV% (low even for a spot-up shooter) is short: Al Jefferson in 2012. That's it. Even if you relax the standards to 8 TOV% the list is still pretty modest, with three seasons each from Al Jefferson/Dirk/Michael Redd, two from Jordan, four from Jamison, and a scattered few others. With an absurdly low turnover rate, his offensive is more palatable. For instance, using a tweaked TS% formula for turnovers -- PTS/(FGA+.44*FTA+TOV) -- he's at 0.968, in terms of points per possession, the same for David Lee if you round to three digits (without rounding Al Jefferson is a little higher.) David Lee is often lauded for his stats, and while his generous passing explains some of the disparity in turnovers, it does highlight the perception of efficiency versus reality.

The Bobcats will definitely be better on offense, thanks mainly to the aforementioned Al, but the defense will still be one of the league's worsts, unless we see a major leap from someone like Kidd-Gilchrist. Lighter minutes from Biyombo and Haywood will sink the defense even further, unless the Bobcats somehow manage to play Biyombo or Haywood next to Al, but the main difference is that they will no longer dedicate huge minutes to BJ Mullens, which helps on both ends. A core of Walker/Henderson/Kidd-Gilchrist is actually pretty promising if you're an optimist, but I'm not sold on the Zeller/Al alliance, and given the general incompetence of the organization I'm erring on the side of terrible. To end on a hopeful note, the LA Clippers were once the long-running joke of the league, thought forever to be the clowns of the NBA and always a disappointment, but a great draft pick in Griffin and a landmark move later with Chris Paul, and their fortune has been completely changed. Those kind of trades are uncommon, but the takeaway here is that it's a possibility. That's all a Bobcat fan can hope for. Drastic, positive change can happen.

Random prediction: Biyombo will turn a corner defensively and display some of that potential, but he'll be so buried on the bench sometimes it won't matter.

Win/loss prediction: 22-61


Boston Celtics

The Celtics blew up the team a year too late, riding the aged Pierce and Garnett into a weak first round loss to a team that wasn't even great. The gamble half a decade ago had worked, in what now must feel like an obvious move, riding defense and team play into a playoff berth every season along with finals appearances and a title. At the time, there was some criticism at mortgaging the future (losing picks and Al Jefferson) for players on the wrong side of 30. Well, the Celtics rode this core for further than they thought they could, and they should be thankful they didn't come up empty.

How Boston's season is viewed is based on what you think of Rondo, the lone remaining star. As I've covered before, he's an overrated offensive player as the Celtics don't seem to miss him when he doesn't play. What's particularly damning is the fact that they haven't even had good back-up point guards, relying on the likes of Eddie House and Nate Robinson and, last season, Avery Bradley. Yet their offense doesn't crumble when he's gone. While Bradley is a defensive force, he can't run an offense, but the Celtics improved when Rondo was sidelined. How did this happen? The ball was distributed more to Pierce and even Garnett; those two were, for various reasons, more effective focal points for their offensive system. One often discussed topic is Rondo's shooting, and even though he shot the ball well from the midrange last season what's important is that teams give him a ton of space and use his defender to jam the rest of the offense. This can kill spacing, and because Rondo is typically too passive in shooting he doesn't make them pay for this approach. The ball is often sticky in his hands, and he's a culprit of "stat-padding" his assists sometimes, looking for stats rather than simply the best possible option. (For an example, he was actively hunting for assists during his double digit streak last season.)

With Rondo
Without Rondo
Season
Games played
Adjusted offensive efficiency
Adjusted defensive efficiency
Games played
Adjusted offensive efficiency
Adjusted defensive efficiency
2008-10
238
111.2
100.8
8
-2.0
0.6
2011
68
108.1
102.1
14
-5.0
-3.2
2012
53
101.9
99.7
13
-0.7
-3.8
2013
38
102.3
103.7
43
2.3
0.2
2008-13
397
104.2
102.4
78
0.0
-1.0
Regular season only. Efficiency is in terms of points scored/allowed per possession. The without Rondo numbers are relative with the with Rondo numbers. For example, -2.0 on offense means the team is 2 points worse without him. Adjusted means it's adjusted for strength of the opponent.

While he was missed during some stretches, the trend has actually been downward and overall he's been a net zero on offense despite a lack of true point guards behind him (on another note he still rates as a plus defender through this method.) Rondo is a unique player in many ways, raising his game in bigger moments, but his passing exploits don't seem relevant to a team without shooters or scorers. I don't feel that his playmaking will "make others better" because we don't have evidence he's done so in the past; for example, he usually rates near the bottom in the league for ratio of at the rim assists to long two-point assists, which are the least valuable. However you view him, he's still on the injury list from the partial tear of the ACL, and the Celtics may decide to give him the Rose treatment with extra rest.

The rest of the cast leave a lot to desire. Since they were in competitive purgatory, not bad enough to get a real draft pick and not good enough to do anything in the playoffs, they were granted with the 13th pick, Kelly Olynyk. Olynyk is a Gonzaga player with the rare and dubious distinction of having a smaller wingspan than his height, which is troubling in basketball. He's skilled and he can shoot, but there are question marks about how he can adjust to the NBA, not surprising given his draft position despite no health or character problems. He may provide some value as a stretch four, but it'll take a lot to save an offense that was bad with Pierce and Garnett last season. Any enthusiasm for him should be tempered; betting that a late lottery pick is an instant success is a good way to lose consistently. As for last year's draft pick, Jared Sullinger continues to have injury problems, and even if he's on the court his defense will erase any benefits he can give from rebounding and scoring.

Jeff Green is a popular pick for a breakout player and for whatever reason the perception of him as a player outweighs the reality. After years of injuries and disappointing seasons, he finally ... became an average player. His PER was 15, league average, his Win Shares per 48 minutes was 0.099, near average, and his xRAPM was -1.4, which was near the league's median. He got to the line more and shot well from outside the arc, where he's normally been inconsistent. However, for a power forward he's a terrible rebounder, and for a wing player he's a bad playmaker. He improved a lot after the all-star break and that's how people view him now, but that only reinforces the point that he's streaky. People now assert that Perkins for Green was a folly, but they've forgotten the multiple years that Green was a terrible player and how he was a poor fit on the Thunder next to Durant while taking minutes away from Ibaka. He's also already 27 years-old, meaning his ceiling as a player as an athletic forward has probably been met.

The only other notable player besides the newcomers is Avery Bradley. He rated as one of the best defensive perimeter players, competing with elite guys like Iguodala and Toy Allen. Via Synergy, for example, he allowed the fewest points per possession against for point guards. It's a myth the Celtics improved when Rondo went down because of their defense; the above table shows it was their offense that improved. However, Bradley is not a good offensive player. He can't shoot or run an offense, meaning his position is unclear. With Pierce, his inability to run an offense was masked because the future Hall-of-Famer took over the playmaking duties; he'll no longer have that advantage and while Rondo's out the offense could be really ugly.

In exchange for the second leading scorer in the storied Celtics history and the defensive force-of-nature, all-star big man, they received Gerald Wallace, MarShon Brooks, and Kris Humphries, as well as some picks. Gerald Wallace at his peak was a destructive player who averaged double digit rebounds one year as a perimeter player, which is rare in the modern game and in the last decade arguably the only other player to do this was Marion, and another season he averaged 2 blocks and 2 steals, only accomplished by Olajuwon and David Robinson (and Dr. J in the ABA.) Unfortunately, age hasn't been a kind foe, and Gerald has crashed offensively, becoming a liability on the court. He might regain some of his offense, and given his comments in the preseason it doesn't appear he's giving up, which jives with his on-court playing style. Otherwise, MarShon Brooks is already 26 and at his best is a bench sparkplug scorer with low efficiency. Kris Humphries will help push the Celtics into an acceptable range with their offensive rebounding. The 2013 Celtics were the second worst offensive rebounding team ever trailing only their 2012 season; three of the bottom five seasons were from the Celtics from a deliberate strategy of eschewing offensive boards for transition defense. Expect them to bounce back with a new coach and Humphries replacing Garnett.

The good news is that the Celtics have ten first round picks in the next five drafts and the finances look rosier, with only problem being Gerald Wallace's ten million a year contract until 2016. Rondo's future with the team is in doubt, but otherwise Boston could be completely reloading and what we're seeing is a transition year. Any expectation of chasing the 8th slot in the playoffs should be squashed since they won't have the defense to counteract the terrible offense. You can't have a great defense with a frontcourt of Bass, Olynyk, Sullinger, and Humphries. On a final brighter note, given their roster and the likelihood of Rondo's departure, they won't have to worry about drafting for a specific position; they'll need upgrades everywhere.

Random prediction: Jeff Green's shooting percentages will go south, 50 to 52 TS%, and he'll look uncomfortable on the floor for long stretches without Rondo.

Win/loss prediction: 24-58


Milwaukee Bucks

Bringing back only four players, the Bucks have entirely revamped their roster while retaining key young guys. Jettisoning master bricklayers Jennings and Ellis, the offense will be more balanced, relying on the mix of guards Knight/Ridnour/May to some extent and leaning more heavily on their role players. Due to the lack of shot creators the role players could be more pressured to score, but it's not like Jennings and Ellis were efficient with their opportunities. Their offense will still be terrible, but the defense should tread water above the average rate largely due to Larry Sanders in the middle.

Every season there are a small number of players who come out of nowhere: sometimes it's just one player, sometimes five. Larry Sanders managed to increase his FG% by 5 points and his defensive rebound rate by 8.5 points. Given his age, 24 last season, his improvement was a rare event. His defensive numbers were amazing, as the Bucks defended like one of the best teams when he was on the court and were below average when he left the court. Given his foul rate and ejections, he was all too often off the court, and his biggest area of improvement now is minutes played. He stands only 6' 10" without shoes, but with his enormous wingspan only a handful of players have a higher standing reach in the league. With his quickness, shotblocking ability, and Boeing-like arms, he can shut down the paint and posting him up doesn't appear to work well either even with his lanky fame: Synergy found opponents averaged 0.7 points per possession in the post against him, and in isolation he ranked 21st in the league at 0.61 points. His 7' 5.5" wingspan can be joined in the frontcourt with John Henson's 7' 5" wingspan, terrorizing opponents, and the exciting prospect Giannis Antetokounmpo brings a frightening 7' 3" wingspan to the SF position.

Giannis will hopefully let Bucks fans forget the ridiculous Redick trade where they lost tantalizing player Tobias Harris. Scrambling to hold an 8 seed for an impossible series versus Miami, Milwaukee decided to trade a prospect to rent Redick, who would probably leave at the end of the season -- and he did. Nevertheless, Giannis appears to be the future at SF, and a year ago he was virtually an unknown player who grew up poor in Greece before shooting up a few inches with a massive growth spurt in his mid-teen's. His body is nearly a carbon copy of Kevin Durant's, and while he may not project the same as a scorer he's a surprisingly intuitive player for someone with little experience with nice ballhandling skills, good timing on his blocks, a nice touch from outside, and a few tantalizing skils like a Dirk step-back, knee-up jumper that's eerily similar to Durant. When international prospects haven't had significant time in an elite league, it's a bigger gamble investing in a player because even scouting is hampered by the fact that you don't know how he responds playing with the best-of-the-best. But when it works, you look smart.

The imports besides Giannis are all lukewarm players who will help maintain the Bucks' tradition of mediocrity. OJ Mayo was highly coveted at one time, but he's settled into being a decent jump shooter who can't handle a bigger load and doesn't offer much else of value. Ridnour is one of the best midrange shooters in the game, but he remains a below average guard -- it's not like he's Dirk Nowitzki out there. Carlos Delfino comes back to Milwaukee, but past 30 he's nothing more than a veteran wing. The Bucks curiously signed Zaza Pachulia, a tough veteran big man instead of devoting more minutes to their young guys, but this was part of their desperate attempt to stay relevant. Gary Neal rounds out the guard slots as another excellent midrange shooter, but his defense was terrible even though he was schooled in San Antonio.

The Bucks crawled into the playoffs last season with 38 wins, though given their point differential they were closer to a 36 win team. Slipping offensively without any big scorers, the Bucks may miss the playoffs as a few eastern conference opponents load up, like Detroit with Josh Smith, Washington with a full year of Beal, Wall, and now Gortat, and the Bulls leaping from mediocrity with Rose back from injury. They'll need another breakout season like the one they had from Larry Sanders, and while all eyes are on Giannis he's only 19 years and may have another season before he has a significant positive impact. Long-limbed and full of youth, Milwaukee's frontcourt is a promise for the future, and here's hoping the front office won't ruin their potential with short-sighted moves.

Random prediction: Giannis has a 4-4-4-4-4 game (points/rebounds/assists/steals/blocks) just one steal and one block short of the rare 5 by 5.

Win/loss prediction: 30-52


Washington Wizards

The Wizards are a popular pick for one of the most improved teams and a playoff appearance because when Wall came back from injury, the team tore off and separated themselves from the muck and mire of the eastern conference cellar-dwellers. People assume Wall is the catalyst here, but it's really the whole lineup -- Wall-Beal-Webster-Nene-Okafor. It was a frightful defensive unit, one of the best in the league, and they actually didn't score all too well. Nene and Okafor are the two underrated cogs here: Okafor is the quintessential rim protector and rebounder; for his career he's around a 25 DReb rate, which is an elite level. Nene is a subtly skilled big man who does well at numerous little things, like boxing out, where his strength makes a post-up a difficult and his quickness alleviates any problems with pick-and-roll's. Unfortunately, Okafor went out with a back injury, and the Wizards responded by trading him for Gortat. He's a much better offensive player who should stay on the court, but he's not the same level as a defender. Without any help on the bench, the frontcourt defense could take a plunge.


Explanation of the above figure: it's on a per 100 possession scale. Typically with team strength you use point differential, how many points you outscore the opponent, but here I'm doing point differential on a 100 possession scale with the gray line. It's a little crowded, but the information is useful. With the black dots you see single games. A moving average is just a weighted average of the past 8 games, where the most recent game is more important.

How important was Okafor? Using the valuable website NBAWOWY.com, when Nene was on the floor without Okafor their offensive rating was 102.4 per 100 possessions and their defensive rating was 103.3. With Okafor and Nene? 104.2 on offense and 99.5 on defense. The previous ratings suggest a below average team in the mid to upper-30's for wins; the latter would be a 50 win team with a good seed for the playoffs. For some hope, when Wall-Beal-Nene were on the court with Okafor on the bench, the offensive rating was 108.0 and the defensive rating 106.4. That's a decent team, but Nene can't play heavy minutes and he typically misses a few games per season. This will mean devoting a large amount of minutes to units without both Nene and Gortat, and it will be difficult to defend unless the Wizards pick up someone else.

An improved John Wall can provide enough production to overcome some of this, but, again, injuries: the chances that he will play nearly every game are low. Whether or not his jump shot improves is a key determinant to his all-star chances and how the season goes, but he's found the perfect backcourt mate in Bradley Beal, who shrugged off a low percentage in college from outside to showcase one of the best shooting strokes for a young player. He wasn't perfect, only shooting 35% from his long two-pointers and under 80% from the line, but he's very young and the combination his three-point shooting and ability to finish at the rim complements Wall. He also seemed to be dependent on Wall, as his TS% dropped from 58.2 to 48.8 whenever Wall stepped of the court.

Outside of the main core, the cast isn't very impressive, which is to be expected for a subpar eastern conference team. Seraphin is the first big man off the bench, and while he's a fine scorer his lack of defense and rebounding will cause a disruption. Martell Webster is a decent 3/D wing player, but he's reached his ceiling after a surprising eight years in the league. Ariza was a breakout player for the Lakers in the playoffs, thanks for an unsustainable jump shooting display, and when he fell back to Earth the next season he disappointed with what was once an underrated player. With his size and athleticism, he's a very good defender and finisher near the rim, and his shooting has finally become acceptable but only in the corners, where for two straight seasons he's been above 40%. The other Trevor, Booker, is a small power forward who relies on doing the little things, and he's accompanied by Al Harrington, who's started his journey as a 30-something non-all-star, which rarely ends well.

If they want to reach 0.500 basketball, they'll need substantial seasons from some of their other young players. Jan Vesely is beginning to look like a lost cause, as he was one of the worst players to receive major minutes. Score one for the Hollinger draft/Euro translator, which saw his projected PER at 10.7 (for his career so far, he's been at 10.1.) Hollinger even specifically said taking him in the 5 to 10 range would be a mistake; the Wizards took him sixth. He can't shoot, throwing up airballs with regularity, including the free throw line, and he's turnover prone without any redeeming qualities to overcome his affliction with putting the ball in the basket. Even with his athleticism and length he's not much of a defender. Otto Porter Jr., however, looks like a better bet -- he's a large wing with few weaknesses and many tools. He's a wiry defender who will add strength with time (he's only 19 years-old) with range, ballhandling abilities, and the ability to draw fouls. He's type of no-nonsense, jack-of-all trades player who'd be an asset to almost every team. But as a teenager, there's a small chance he's going to contribute enough to the Wizards to push them over the top this season.

The late season surge wasn't because of one player; it was because their key players were finally healthy. Unfortunately, injuries have already derailed the season. The Wizards last season were the worst offensive team in the league, and only won 30 plus games because of their 5th ranked defense. Okafor's absence in the middle will prevent them from another surge of world-class defense, but Gortat at this point in their careers is just as good overall and he can play more minutes. But this will only be an incremental improvement compared to last season, and without major leaps made by the young players on the team the playoffs aren't guaranteed. Given the shallow depth in the east, it's definitely possible they make the playoffs, but they'll still be sub-0.500 and used only as fodder for a team like the Heat. John Wall, this is what your extension has brought.

Random prediction: John Wall will be third in the league in assists per game, as Westbrook will drop-off due to his injury, Vasquez has to deal with the antics of the Kings, and Holiday and Deron will both have the ball in their hands less due to new teammates.

Win/loss prediction: 36-46


Cleveland Cavaliers

This roster is secretly a mess, and I don't think most people realize it because they're distracted by Kyrie Irving. Alongside the preternaturally gifted scorer, they have a shoot-first undersized 2 who can't shoot, and to complement the two non-defenders they drafted, number one overall, a 6' 7", bulky power forward to use on the perimeter; but if he can make it as a power forward, then he can take time away from promising fellow Canadian Tristan Thompson. The Cavaliers depend on the feisty Varejao, whose defense is sorely needed, but he has trouble playing half the season. But it's okay because they brought in Andrew Bynum, who missed all of last season because of "injury" and is one of the few players who is far more likely to miss more games then Varejao.

Not to say that the Bynum contract itself was bad -- with the provisions it has, they're protected well. And if he plays like it's 2012 again, they have a major weapon and could be deadly in the playoffs.  But I also wouldn't be surprised if he only plays 500 minutes -- or not at all. And to me, their draft pick screams "Bargnani" bust potential, but the fans are hoping for Paul Millsap (though the real Paul Millsap was uncourted in free agency and signed a small contract.) It's unclear what will happen because his stats including his measurements were underwhelming for a true star, and he's old for a one-year college player. In any case, he doesn't project to be a great defensive player, and that's what they need.

Even with Mike Brown returning, the number one pick (again), Bynum, and a few other interesting pieces, the team is still centered on Kyrie Irving, the Australian/American with the best set of handles outside of Chris Paul and with an unusually great jump shot for a young point guard. In fact, that might be the best point of reference: a uniquely quick and skilled point guard, though he lacks the passing of the Clippers guard. But the question is, if Irving is a true superstar, why have the Cavaliers been so terrible? Is he worse than his numbers suggest?

When he was on the court, the Cavaliers only had an offensive rating of 105.7, which was slightly below the league average (via b-ref.) Off the court? 103.0. And defensively, they were 1.7 points per 100 possessions worse with him on the court. Using 82games.com, you can check the offensive (and defensive) rating of each five-man lineup. The most common lineup was Irving-Waiters-Gee-Thompson-Zellers. The sixth most common was the same but with Shaun Livingston instead of Irving. What was the drop-off? A modest 4 points per 100 possessions -- not what you'd expect swapping out an all-star scoring point guard with a backup point who's had devastating injuries. But what if Irving needed a real center, like Varejao in the middle? Now the offense is at 114 points per 100 possessions, a scorching rate that's nearly league-leading, and it was the second most used lineup. That same lineup with Pargo instead of Irving? 110 points per 100 possessions -- again, only a modest drop-off, Pargo was lucky to get minutes with an NBA team, and it was the seventh most used lineup, so it's not a case of small sample size. For all the love of Irving, the team's struggles have more to do than Irving getting more minutes, and if you complain the lineups are used in different circumstances, line versus weak bench units, adjusted methods like RAPM don't paint a rosier picture: one version of "pure" ridge-regression adjusted plus/minus had him at +3.1 on offense and +1.7 overall, while another method mixing in box score stats had him at +3.0 on offense and +0.9 overall. Simply put, those are not superstar numbers, and if it continues to be a significant negative on defense he'll have to be an otherworldly offensive player like Harden or Curry who rate +5 to +6 on offense, depending on the method applied. For whatever reason, his fantastic numbers and jaw-dropping plays have not translated to the team level yet.

Minutes rank
Lineup
Minutes
Off. rating
1
Irving-Waiters-Gee-Thompson-Zeller
421.6
104
6
Livingston-Waiters-Gee-Thompson-Zeller
99.2
100
1
Irving-Waiters-Gee-Thompson-Zeller
421.6
104
2
Irving-Waiters-Gee-Thompson-Varejao
189.5
114
7
Pargo-Waiters-Gee-Thompson-Varejao
99.0
110
82games.com, 2013

But Irving is unbelievable in the fourth quarter and especially in clutch situations. Per NBA.com, he led the league in shots taken in the last five minutes of a close game per game. He shot an impressive 46.7% from the field, and including his free throws and three-pointers it was a scalding hot 71 TS%. And this is no fluke: in his rookie season he shot 54.4% from the field. The fact that he's able to take more shots on better efficiency, on a team without much help, is remarkable.

The team is hoping to land in the playoffs and win more than half its games, but they were fairly awful last year and what has changed? Their major additions are Bynum (good luck) and a shoot-first, midrange specialist bench player in Jack (appropriate name.) They'll hope for more minutes from Varejao, who hasn't played more than 40 games in four years, and Irving who's been limited by injuries every year including college. They have a few young players who could break out, but there's no counting on that. So you're banking on a productive year from Bynum, good health from a couple guys without that track record, young guys who have yet to prove themselves, and a boost from a former coach who was fired last year after five games. Even if they get lucky with a large number of games from Bynum and Varejao, they play the same position, and it doesn't help the fact that they team lacks outside shooting and quality wing players, depending on guys like Alonzo Gee. They could very well squeak into the playoffs, but this team has a long way to go for any serious noise in the east.

Random prediction: Bynum and Varejao will get injured in the same week, instantly killing their center depth.

Win/loss prediction: 37-45


Detroit Pistons

Just call the team the spacing experiment. Detroit wants to start Monroe and Drummond, even though the two are more centers than power forwards, and Monroe's jump shot has been inaccurate: from 16-23 feet he was 30% last season. So of course, joining them at small forward will be Josh Smith, who managed to best him with 33%. Floor stretching isn't Smith's strength, as plenty of fans and media members have pleaded to him to abandon his three-point shot. To operate this offense they brought in Brandon Jennings, a guard who's fine with throwing up ill-advised jumpers even if he rarely cracks 40% from most zones. The likely starting shooting guard Stuckey isn't any better than Jennings. Luckily, they do have some half-decent shooters on the bench like Singler and Billups and perhaps Caldwell-Pope, but they probably don't want to rely on the 37 year-old Billups or a rookie.

Optimists will point to the Grizzlies as an example of an elite team without shooting, but that's misleading. For one, the Grizzlies were elite because of their defense, which was at an absurd level: normally you only see one team a season with a defense +5.5 points per 100 possessions better than average. The Pistons were far below average on defense last season and there's virtually no way this new group can even approach Memphis' stifling defense. Memphis, to their credit, recognized the glaring lack of outside shooting and targeted Mike Miller. Even with Randolph and Gasol on offense they were barely above average on offense, and last season they had significantly better outside shooting. Prince is a better shooter than Smith, Conley's a reliable three-point shooter, Pondexter shot 40%, and both Gasol and Randolph are good midrange shooters. Marc Gasol shot 48.5% from 16 to 23 feet, one of the best conversion rates in the league. So yes, you can be a great team with poor outside shooting, but you need an elite defense to overcome the likely mediocre offense, unless you have some uniquely great players on your roster. You can point to Denver's success without outside shooting, but they had a great system, attacked in transition, and still had some good shooters like Lawson. Detroit doesn't have the same personnel. They will not be special on defense, and their offense wouldn't be great anyway, regardless of spacing.

The power of spacing is why three-point records continue to be broken and teams, especially smart teams, have focused on outside shots. People were worried about the Monroe-Drummond combination, much less adding Smith to the starting lineup. Last season, the team was hesitating on the big man pairing, partly due to defense as Monroe wasn't quick enough to guard the perimeter. Using NBAWOWY.com, when Knight-Prince-Monroe-Drummond were on the court, they scored 0.951 points per possession. When Knight-Prince-Monroe were on the court with Drummond on the bench? 1.020 points per possession. That's bigger than the difference between the offense of the Raptors and the Thunder. Obviously, that's just one example, but it's clear on a general level having so few shooters on the court can dampen an offense.

On the bright side, Andre Drummond had a jaw-dropping season. After a disappointing season in college where people questioned his effort and skill level, Drummond posted some of the best per minute stats for big men. Not just rookie big men: everyone. He had a 22.1 PER in college and posted a 21.6 PER in the pros. That's not supposed to happen. The strength of competition differential between the NCAA and the NBA is astronomical. Drummond was one of the few men of his size and strength in college; in the NBA every team has someone his size and many have guys who are at least nearly as athletic. Yet somehow his FG% increased by 7 points. Few big men in history have offered his combination of finishing skills, rebounding, and shot-blocking at his age.

Name...................
Season
Age
Mins
FTA/
FGA
FG%
TS%
USG%
BLK%
TRB%
Sim. Score
Andre Drummond
2013
19
1243
0.46
60.8
57.8
17.2
6.1
21.2
0
Amir Johnson
2008
20
764
0.33
55.8
58.4
14.1
8.5
18.4
0.084
Andrew Bynum
2008
20
1008
0.40
63.6
65.9
17.5
5.1
19.6
0.052
Shaquille O'Neal
1993
20
3071
0.55
56.2
58.4
27.0
5.7
20.6
0.096
Greg Oden
2009
21
1314
0.64
56.4
59.9
19.3
4.2
20.0
0.085
Dwight Howard
2007
21
3023
0.76
60.3
61.9
22.7
4.2
20.5
0.150
Shaquille O'Neal
1994
21
3224
0.53
59.9
60.5
29.0
4.5
18.7
0.134
Seasons where Mins > 500, Age < 22, TS% > 56, BLK% > 4, TRB% > 18

If you open the gates to 22 years old, then the following players show up: Oden again, Kosta Koufos, , LaSalle Thompson, Shawn Kemp, Tyson Chandler, Zydrunas Ilguaskas, Hakeem Olajuwon, and Dwight Howard again. But Drummond is the youngest person on the list and production and skill accelerate quickly at such a young age. Sim. Score is a similarity score comparing each player to Drummond by standardizing FTA/FGA, FG%, USG%, BLK%, STL%, and TRB% to the same scale and calculating the sum of squared differences. Basically, it's like all the categories are treated the same, and the larger the difference the more there's a penalty (since it's squared.) Andrew Bynum was extremely similar in 2008 because he scored with a high FG% and had great rebound and block rates. Drummond rated better as a rebounder and shotblocker, and was a year younger, but their usage rates were, oddly enough, very similar. Bynum, however, had a low steal rate, while Drummond's was easily the highest out of this group, suggesting elite quickness for his size. By the way, I also ran a Sim. Score including free throw percentage, but it wasn't as useful because of how terrible Drummond was from the line (no one is similar.) But it's worth noting the three best scores change to Shaq (0.145), Oden (0.156), and Bynum (0.157.)

Obviously, Drummond has high potential, but the rest of the roster is misshapen with only a few intriguing pieces. Greg Monroe is one of the best offensive big men in the league, scoring at a decent clip to go along with great passing. He's a copy of Josh Smith in some ways, where he can pass very well but shoots jumpers too often at a low efficiency. Monroe assisted 18.6% of his teammates field goals last season. Since he operates from the high post usually near the elbow, he could create a very effective two-man game with Drummond throwing lobs. The only other player who assisted and rebounded like Monroe last season was Joakim Noah. The other major addition was Brandon Jennings, who was swapped for Brandon Knight. He's an aggressive shooter who will pull up for bad shots and he's a poor finisher inside, but he may be more inspired to pass this season given the attractive targets; with the Bucks he was sometimes forced into the shot creator role, whenever Ellis wasn't clanking a shot.

The rest of the team consists of odd pieces and the type of guys you'd see on a mediocre team. Stuckey is already 26 and never filled his potential as a scorer. Singler is an intriguing shooter but his ceiling is pretty low. Billups coming back to Detroit is, using the best possible comparison, a victory lap before retirement. Though he could be buried with Josh Smith and Monroe using more forward minutes, Jonas Jerebko is a tall Swedish shooter who will be called upon to space the floor. It may not be this season, but Caldwell-Pope looks like a 3/D player who fits well on the roster next to Jennings as a plus defender with range. His development will be key to the big lineups working because the other SG, Stuckey, is a slasher without a jumper and you don't want to rely on Billups.

Make no mistake: the Pistons will be a lot better due to the development of some of their young guys, giving Drummond more minutes, and the addition of Josh Smith. They won only 29 games last season, and I see them approaching 0.500 ball and maybe a playoff spot. That's a large improvement, even with all the spacing problems. It was an expensive shopping season with the best case scenario taking one game from Miami in the first round, where Drummond eats through the likes of Bosh and Chris Anderson while briefly meeting past doppelganger Oden, but it takes some gambling to move up in the league.

Random prediction: the Pistons find the most success with Smith-Drummond and Monroe on the bench, and Greg Monroe will lose his starting spot late in the season as the Pistons try to make the playoffs.

Win/loss prediction: 39-43


Atlanta Hawks

After nearly a decade, the Josh Smith era is over. Danny Ferry has completely restructured the team, bringing financial flexibility along with some smarter contracts. This is more of a rebuilding project like Houston undertook, cycling through role players while remaining competitive enough to fight for the playoffs. As talented as Josh Smith was, he was strongly criticized for his shot selection. Scoring is weighed too heavily in NBA evaluations, so most people emphasized his weaknesses and ignored his strengths like how he's one of the best passers and ballhandlers at the power forward position; he's a great defender who can play either forward position and protect the rim. But he relies too much on his midrange jumper when he has one of the worst jump shots in the game.

It's strange how people criticized Josh Smith for his ill-advised jump shots but when they replace him with a quality player with a nice jump shot they're seen as a team in decline. The Hawks were almost exactly an average team, a point differential close to zero, and they've managed to swap out a player who relied on his athleticism and a broken jumper with a highly skilled, jack-of-all trades player who provides the same level of production but with a contract under 10 million dollars -- DeMar DeRozan makes the same amount of money, and Millsap will make less money than guys like Ty Lawson. Millsap is perhaps the most underrated player in the league. He grades out well in every metric. His PER is around 20, year in and year out, which is on the borderline of being an all-star player. His adjusted plus/minus numbers are sterling too; by xRAPM he was the seventh best player last season at +5.4, and a "pure" plus/minus only metric graded him eleventh at +4.5. He's not a rim protector, but he picks up a lot of steals, rebounds well, and is fundamentally sound where his quickness can guard some perimeter players. He shoots from virtually everywhere, and he also appears to be unleashing a three-point shot for this season.

Millsap will join his competition for most underrated player in the frontcourt next to Al Horford. Like Millsap, Horford is a well-rounded player, jack-of-all trades, but he is master at something: midrange jumpers. Few people know this, but Horford has an elite midrange jumper and he annually rates out high in percentage and volume. The graph below shows the top 31 shooters by attempts (via Hoopdata.com) from the past three regular seasons (31 because Horford and Griffin were tied with attempts.) Horford belongs in the same class as guys like David West and Chris Bosh. In fact, besides Dirk Nowitzki, Horford has the highest percentage. And at the bottom of the graph you can see Josh Smith, who was one of the worst high volume midrange shooters in the league. Millsap, by the way, grades out decently in the Aldridge/Jefferson/Boozer group, and replacing Smith's jump shots could add roughly 30 points over the course of the season, translating to another one win added just through his jumper accuracy.

Danny Ferry has turned over the roster masterfully and is trying the impossible challenge of rebuilding while remaining in playoff contention, something the Rockets recently accomplished. They have some great core players even elite teams would want to have, like Horford and Millsap, but also the sharpshooter Korver and even the young point guard Jeff Teague. In fact, with Teague, they have a small wealth of young talent, including the German who has impressed people over the summer Schroeder and intriguing talent Lucas "Bebe" Nogueira. To round out the depth chart, they picked up Elton Brand and Gustavo Ayon. The former still rates well as a defender even though he's hitting his mid-30's. He's very strong and has a condor-like wingspan; he has 1700 blocks for his career, amazingly enough, and has a decent shot at 2000. Ayon is a productive end of the bench big man with a nice touch inside and a surprising passing ability. If Lou Williams recovers from the ACL injury, the team should have enough depth to compete over the 82 game season.

The Hawks will lose some firepower on the defensive end without Josh Smith swatting everything, but Millsap's all-around play will be an upgrade on offense; overall, there's hardly a change in their overall talent, excluding a breakout season from one of their young players. They should maintain their 0.500 winning record, give or take a couple wins, before bowing out in the playoffs versus someone like Chicago. But Hawks fans are probably still happy from being free of Joe Johnson's albatross of a contract and may fall in love with the young talent.

Random prediction: Dennis Schroeder has a magnificent season and ends up on the NBA All-Rookie First Team.

Win/loss prediction: 41-41


Toronto Raptors

I was befuddled at Toronto's projected standing, and I went line-by-line through the projection and at every player. I'm opening their season preview with a player no one would expect: Alan Anderson somehow played nearly 1500 minutes. He graded out as a terrible player across the board, one of the worst to receive so much playing time especially for a team close to the playoff race. As criticized as Rudy Gay is, he's a much better option at small forward. Steve Novak is one-dimensional, but it's one more dimension than Alan Anderson had. Anderson actually hoisted a lot of outside shots but clanked two-thirds of them; Novak made 50% more total, but only had 23% more attempts. Another notable loss is Andrea Bargnani, a tall shooter who can't shoot or do big man things. Taking his minutes will be Tyler Hansbrough and Jonas Valanciunas, and to an extent Amir Johnson. While Hansbrough isn't a major plus, he still rates better than Bargnani by many metrics, and Jonas looks ready to make a big leap forward and was better at both ends of the court even as a rookie. Replacing Alan Anderson and Bargnani even with average players: that leads to an improved team, hands down.

The question heading into the season concerns the eyesight of their leading scoring, which after Alan Anderson is the second weirdest way to introduce a team. For some reason, Rudy Gay never realized he was blind, which may help him with his long-range jumper. One under-reported note is that Gay's shooting percentages climbed back to his historical averages once he left Memphis, even though his usage rate ballooned to 29%. On a team with a poor offense, there is value in having a player who can hit a 52 TS% mark without many turnovers or assisted shots. That's close to Toronto's average, and it's close to the league average. While it would be better to dump it to Amir Johnson inside where he's at 60.6 TS% overall, that is not an automatic play that happens every time. Amir Johnson is more opportunistic and relies on others to set him up more than Rudy Gay. At some point you're going to need play/shot creation. Of course, it would be better if you have a wizard of a point guard to slash and kick and set everyone up, but Lowry isn't Chris Paul. That's not to say Rudy Gay is an all-star caliber player, but he's not a huge negative either.

Perhaps to make Rudy Gay look better in comparison, however, is DeMar DeRozan, a smooth dunker with a terrible shot. Referring back to the 16-23 foot shot chart in the Hawks' section, DeRozan takes an absurd amount of long two-pointers, and makes Rudy Gay look shy by comparison. At least both are excellent at the rim, normally finishing around 65%. Gay actually splits his shots evenly by volume through five zones: at the rim, 3 to 9 feet, 10 to 15 feet, 16 to 23 feet, and three-pointers. By contrast, DeRozan loads up more 16 to 23 foot jumpers with half as many three-pointers. He's a poor shooter behind the arc, but he's a poor shooter inside too, and that extra point behind the arc makes it a worthy trade. Surprisingly, he's taken the fourth most shots from that distance over the past three seasons, due partly to his ironman status, but on a per minute rate he's up there with Kobe, Aldridge, and Carmelo. Besides the dunking, the most positive thing I can say about him is that he can play huge minutes, being one of only five players over 3000 last season and playing a ridiculous 60 minutes in a triple overtime game. The dynamic bricklaying duo of Gay and DeRozan will cause some hair-pulling from fans. To add to the fun, don't be surprised if Tyler Hansbrough raises his usage by a couple points to around 22/23; his career TS% fits in well with those two.


Toronto manged a league average TS% despite poor leading shot creators. At least Anderson and Bargnani are gone, and their shots will probably be distributed first to Jonas and then to the rest of the team. Gay and DeRozan will remain the focal points of the offense to the dismay of many fans, but the rest of their opinions are efficienct from Amir Johnson to Lowry to Novak and Valanciunas' torrid 62 TS% mark, thanks to strong shooting inside coupled with nearly 80% from the foul line. And we're just two years removed from Novak's 68 TS% season in New York where he took (an unofficial) record number of three-pointers per minute, for players with 1000 minutes or more, yet made 47% of them. Curry is getting all the three-point love, but we can't forget Novak's season. Curry, by the way, only took 7.3 three's per 37 minutes last season, while Novak took 10 in 2012.

Rank
Player.....................
Season...
Team
Minutes
FGA/36 mins
3P%
3PA/36 mins
1
Steve Novak
2011-12
NYK
1020
11.9
47.2
10.0
2
J.R. Smith
2007-08
DEN
1421
17.1
40.3
9.9
3
J.R. Smith
2006-07
DEN
1471
15.8
39.0
9.3
4
Dee Brown
1998-99
TOR
1377
12.9
38.7
9.1
5
Carlos Delfino
2012-13
HOU
1689
13.2
37.5
9.0
6
Steve Novak
2008-09
LAC
1161
12.3
41.6
8.9
7
Michael Adams
1990-91
DEN
2346
21.8
29.6
8.7
8
C.J. Miles
2012-13
CLE
1364
16.2
38.4
8.7
9
Ryan Anderson
2010-11
ORL
1424
13.3
39.3
8.6
10
George McCloud
1995-96
DAL
2846
16.2
37.9
8.6
11
Dennis Scott
1994-95
ORL
1499
15.5
42.6
8.5
12
Charlie Villanueva
2012-13
DET
1092
15.3
34.7
8.5
13
Eddie House
2008-09
BOS
1479
13.6
44.4
8.3
14
Vernon Maxwell
1998-99
SAC
1007
15.1
34.6
8.3
15
Brooks Thompson
1996-97
TOT
1055
13.9
39.8
8.3
16
John Starks
1994-95
NYK
2725
14.0
35.5
8.1
17
Baron Davis
2004-05
NOH/GSW
1581
17.1
33.3
8.1
18
J.R. Smith
2009-10
DEN
2082
17.9
33.8
8.1
19
Ryan Anderson
2012-13
NOH
2503
16.0
38.2
8.0
20
Quentin Richardson
2004-05
PHO
2839
13.3
35.8
8.0
21
Eddie House
2005-06
PHO
1421
19.2
38.9
8.0
*Players with at least 1000 minutes and 8 three-point attempts per 36 minutes

Speaking of efficient players, Amir Johnson was quietly the best player on the Raptors, and the Raptors were a vastly better team with him on the court. They were 8.2 points better per 100 possessions on offense and 7.0 points better on defense. And he did this while coming off the bench for half the year and held up his numbers before and after the Rudy trade. Since Amir Johnson and Bargnani shared PF minutes, is there an interaction here we're missing? And how will the Raptors do with this group of players? The projected starting lineup of Lowry-DeRozan-Gay-Johnson-Jonas scored 101.1 points per 100 possessions (via NBAWOWY) last season and allowed 94.9 in 343 minutes. That same lineup without Johnson and when Bargnani was also on the bench? 97.3 on offense and 108.5 on defense: yikes. It was only 104 minutes, and they were smallball lineups, but they should have scored at a higher rate and the same lineup with Ed Davis instead of Rudy Gay wasn't better. When Rudy Gay and Amir Johnson shared the court in 744 minutes without Bargnani, the team scored 109.0 points per 100 possessions and allowed 100.4. Yet when Rudy Gay was on the court without Amir Johnson and Bargnani the numbers changed to 99.9 on offense and 109.7 on defense, which would translate to one of the worst teams in the league. Whatever the combination, Amir Johnson is the magic ingredient.

How is this possible? He's an extremely active and long defender who averages 1.3 steals, 1.7 blocks, and 4.7 fouls per 36 minutes. Believe it or not, but he's limited his foul rate and finally been able to play bigger minutes. The Raptors defended vastly better with him on the court due to his quickness and leaping ability. On offense, he's a great finisher at the rim, nearly averaging 70% inside. With a decent touch on the line fouling him isn't an option either, which is rare for an energy big man. When he's not wreaking havoc inside, he's setting up for a jump shot near the three-point line. His ability to crash the boards, finish above the rim, and stretch the floor was the likely combination that set his value so high for the Raptors.

Free of Bargnani and with a smart decision maker in Ujiri, the clouds are parting and the future looks bright. Jonas Valaciunas is displaying promise and they're in the hunt for the playoffs. The damage was already done with the extension for DeRozan who may never fulfill his potential and the trade for Rudy Gay, but Uriji may not done yet making moves. A complete rebuilding project might happen, dumping one of their starting perimeter players, but a combination of Amir Johnson and Jonas Valaciunas should still be able to rack up a few wins. And, hey, maybe once Rudy's eyesight is fixed he'll be able to see the rim.

Random prediction: free of a couple deficient long-range shooters, the Raptors move up from 26th on three-point percentage to sixth overall, largely due to Novak and career seasons beyond the arc from Lowry and Gay.

Win/loss prediction: 42-40


New York Knicks

Over the past couple seasons, the Knicks have found a bit of success with vastly different teams despite sharing a lot of the same personnel. The 2012 Knicks were an excellent defensive team led by defensive player of the year Tyson Chandler and caused a large number of turnovers for the opposite side. But on offense they were near league average, and were disrupted by a lot of turnovers of their own. They deployed a few defensive specialists like Shumpert and a little-known point guard Jeremy Lin who had turnover problems. In 2013, however, they experimented with Carmelo Anthony at power forward while Amare was out with injuries. Tyson Chandler battled injury problems of his own and his defense was hampered, destroying the Knicks on that end of the court: besides a limited, aging Kenyon Martin who didn't play the entire season, they had no other good interior defenders. Luckily, this fall was countered by an absurd love affair of three-pointers, mainly by JR Smith, Steve Novak, and Carmelo Anthony, enjoying the freedom at PF with arguably his best offensive season. Carmelo even had a career year behind the arc. The end result? A league average defense but an elite offense just behind the Thunder and Heat.

Unfortunately, with Chandler limping through the playoffs and JR Smith affected by a knee injury, they bowed out to the Pacers, who were the stronger team anyway. What did they learn? How can the Knicks and a Carmelo-led team contend for a title? The 2012 incarnation made use of Carmelo's superior isolation skills: since he needs little help in setting up, besides some spacing, you can surround him with defensive players and combine an elite defense with a decent-enough offense. As is the case with most defensive teams, the leading scorer gets all the credit, a tradition that includes Iverson with the 76ers, Rose with the Bulls, and Isiah Thomas with the Pistons, but that central piece on offense is vital to the operation, overrated or not. But Tyson Chandler is unlikely to regain his award-winning form, and the Knicks have added Bargnani, one of the worst defensive big men in the game, to their frontcourt. With Amare's onerous contract begging him to receive more playing time, the frontcourt rotation will have too many minutes devoted to terrible defenders, and Anthony playing PF won't help.

Offensively, they may lose a bit of their mojo. After setting an NBA record with three-pointers, they've lost their best three-point shooter in Novak, and two other good shooters in Kidd and Chris Copeland. The major additions, Metta World Peace and Bargnani, are below average outside shooters. Beno Udrih is a good shooter, but mainly from inside the arc. By setting Carmelo as a PF and surrounding him with high volume, quality shooters, the Knicks spaced the floor, and let a dynamic scorer have more freedom against less agile defenders. With some shooting gone, their offense will help, and things go could south quickly if JR Smith doesn't return to form after a major injury. Bargnani appears to fill some role as a long-range bomber, but his three-point stroke has been off for a few seasons and he's never been particularly great. Habits are difficult to change, and he's turned into more of a midrange shooter who doesn't shoot very well.

Going back to World Peace, that's the sort of move that makes sense for the team: he's strong enough to guard post players but can still guard small forwards effectively, making him an ideal companion for Carmelo Anthony, and while his outside shooting isn't very accurate if he operates as a stretch 4, relatively speaking, he's a better shooter. But with the other major addition it appears there will be less smallball, as the Knicks plan on starting the Italian seven-footer. So the Knicks decided to abandon the offensive approach that led to success last season, and likely won't have the defensive firepower to emulate 2012.

Actually, the roster does resemble two distinct teams mashed together. The defensive Knicks include Felton, only decent but better than the 35 year-old backup point guard on defense; Shumpert, who only played half a season recovering from a knee injury; the incoming Metta World Peace at both forward spots; Carmelo the lone scorer; Kenyon Martin off the bench; and DPOTY Chandler. The offensive Knicks are Pablo Prigioni, a pass-first point guard with a nice shot; JR Smith the gunner; Beno Udrih the midrange specialist; Carmelo Anthony as smallball power forward; Amare Stoudemire as the bench post-scorer; and Bargnani the jump-shooting big man. For the Knicks to find success, they have to merge the two sides of the roster. But I'm dubious on their ability to accomplish that.

Paying huge stacks of cash for a duo of Amare and Bargnani to get injured and play no defense is not the most prudent way to spend capital on wins. Keeping Carmelo is expensive, and I don't believe New York can construct the kind of successful cast to surround him. They've gotten lucky with some surprises recently, stumbling into a devastating offensive attack, and they'll need that luck again. If Tyson Chandler regains his form and the Knicks ignore Amare/Bargnani as much as possible, using World Peace or Anthony at PF, and Shumpert fulfills his promise as a wing stopper, this could be a very good team. But it appears the Knicks are heading in a different direction, unaware of what exactly makes them a good team.

Random prediction: when Bargnani and Amare go down with injuries halfway through the season, the Knicks find a lot of success with World Peace-Carmelo as the forwards.

Win/loss prediction: 48-34


Chicago Bulls

The future is not the favorite version of your past. Many fans, especially Bulls fans, assume that with Rose the team will be back to 60 wins and competing for the most wins in the league. While a lack of Rose destroyed their offense, it shouldn't decimate their defense especially with such a hard-driving coach. Yet they've slipped from +7 on defense in 2011 and +6.3 in 2012, the best and second best defenses respectively, to a mere sixth in 2013 with +2.7. What changed? Rose is a better option at point guard than Nate Robinson for defense, but the difference can't be all there. Noah had a career season, and Butler emerged as a defensive force.

The secret to the Bulls dominance on defense was their bench mob: a combination of Gibson and Asik, joined often by Brewer, obliterated the opponent's offense. But Houston used a clever poison-pill contract and Asik left, replaced off the bench by Mohammed. When Asik was on the court the defense allowed 93.2 points per 100 possessions and 92.9 points in 2011 and 2012, respectively. With Mohammed the defense allowed 102.0 points. The Bulls had the luxury of two defensive monsters in the middle, which looked great for regular season win totals because the Bulls didn't miss a beat on defense when Noah was out. This also might explain why they didn't look so mighty in the playoffs, by the way, because benches are less important then. Also, the league appears to be catching onto the Chicago defensive philosophy, as the Grizzlies and Pacers have performed exceedingly well in limiting easy shots and stifling opponents. The defense will improve, but that's because, essentially, the 5' 8" Nate Robinson and Richard Hamilton's minutes will be replaced by Rose. And any improvement by Butler would be mitigated by the continuing decline of Boozer and the possibility of one from Deng and Noah, who have already both likely peaked. There are already injury concerns form Noah.

Obviously, the offense will improve significantly. Although Nate Robinson did an under-appreciated job at initiating the offense, just a hair behind Boozer for the highest usage rate with a league average TS% while hoisting up a large number of three's and assisting on over 30% of his teammates' field goals, Rose replacing the minutes from Nate and Richard Hamilton is a huge upgrade on offense. Nate Robinson operated like Rose-lite, hitting a TS%  of 54, which is near Rose's career average. But his usage was far from Rose's: 25 compared to 30 to 32, and the assist rate also paled in comparison: 30% compared to roughly 40%. Basically, Rose will take more of the scoring burden from the rest of his teammates, like Boozer who slumped last year, and he'll make things easier for the role players more suited to playing off of him. If that all sounds obvious, then good: it's a pretty normal order of operations for a team.

Then why won't this team compete for the best record, barring something unforeseen? While Rose is a major upgrade on offense, they had a terrible offense and have a lot of ground to make up. Their offense was as bad as their defense was good, meaning they were basically a league average team. Their point differential was +0.3. So if you assume the Bulls will be a +7 team overall this season, that means you think Rose is worth +7 overall and nearly 20 wins for the average team. That would be one of the best individual seasons in NBA history. When you hear about how many wins players and worth and the various metrics that calculate the estimate, they typically use a baseline of "replacement level player," which is stolen from baseball analytics. Basically, you compare the player to what you would get from a terrible NBA player you could get anywhere for cheap. The problem is that an addition to a team, like Rose, isn't taking away only replacement level minutes. It's pretty close for a starter playing huge minutes, but Nate Robinson wasn't replacement level on offense.

Let's get to another important point, one of the most contentious long-running arguments in the league: Rose's MVP and his status as MVP-level. Winning an MVP doesn't make you the best player in the league. There are plenty of examples of dubious winners. Iverson won an award over Shaq and Duncan in a similar case: a defensive team with a lone scorer getting an inordinate amount of credit for carrying the team. And winning an MVP at an early age doesn't necessarily mean you're on track for an all-time, amazing career -- Hall of Fame track, sure, but not on the level as guys like LeBron. For instance, Wes Unseld won an MVP as a rookie and the same age as Rose, but he didn't turn into Kareem Abdul-Jabbar. His team had a surprisingly good record as they led the league in wins, the MVP was premature, and it wasn't the best season anyway in terms of candidates. Does that sound familiar? The MVP is a media award, and stories are as important as anything. The Bulls jumped out to a surprising record, and in a year when LeBron wasn't in vogue people attached the domination of the Bulls to Rose, who was on the same team that won significantly fewer games the year before. But there were a number of changes to explain their sudden defensive prowess -- and they improved on offense by a huge amount too. Vinny Del Negro was replaced by Thibodeau, while Carlos Boozer, Kyle Korver, Ronnie Brewer, CJ Watson, and Omer Asik joined the team. So they rolled over the new season as a vastly different team with an entirely different philosophy, aided by Rose's development as a player as well as all the new pieces. In light of the context provided the Bulls season isn't a marker for a dominating, MVP-level individual accomplishment but one of team play and defensive discipline with an All-NBA scoring point guard to hold down the offense. Saying that Rose didn't deserve the MVP is in no way negating his value; he's still a great player.

If you were to say Westbrook deserved the MVP for one of his seasons, you'd be laughed out of the room. But the numbers suggest he's been eerily similar to Rose. Both are high usage point guards who hand out assists decently well with a league average TS%. A high usage point guard who still manages to hand out assists is a relatively new trend. In fact, the first season with usage above 30 and an assist percentage above 35 (basically, the player is in control of a huge portion of the team's offense) was in 2004. If you relax the standards to 28 and 33, John Wall and Kyrie Irving sneak in, and they have a great chance at breaking the 30/35 marks soon; and Jordan in 1989 shows up as the first season in NBA history. However, the era before turnovers were recorded were not included in the b-ref search, and assists were actually given out less often in the past because the stat-trackers were stricter in the interpretation of an assist. Jerry West and Oscar Robertson are the two forerunners of this archetype. Rose and Westbrook are carrying that torch, and their numbers are very similar. Both have two seasons that qualify and they're the same age. Westbrook ranges from 31.6 usage to 32.8. Rose? 30.5 to 32.2. Their assist rates are the same too: 38.4 and 42.7 from Westbrook while Rose's rates go from 38.7 to 40.3. Even their TS percentages are similar. Although Rose has less help on the Bulls, Westbrook is the leading shot creator and penetrator for a historically great offense. To say that Westbrook isn't close to Rose's level on offense would be unfounded, and since Rose didn't win an MVP for his offense then we'd either have to concede Westbrook is worthy of an MVP or maybe Rose's award was given out prematurely.

Player....................
Season
Age
Team
Mins.
FTA/
FGA
USG%
AST%
TS%
Team ORTG
LeBron James
2013
28
MIA
2877
0.395
30.2
36.4
64.0
6.4
Russell Westbrook
2013
24
OKC
2861
0.375
32.8
38.4
53.2
6.5
Deron Williams
2012
27
NJN
1999
0.317
30.1
46.6
52.7
-1.7
Derrick Rose
2012
23
CHI
1375
0.344
30.5
40.3
53.2
2.8
Derrick Rose
2011
22
CHI
3026
0.348
32.2
38.7
55.0
1.0
Russell Westbrook
2011
22
OKC
2847
0.454
31.6
42.7
53.8
3.9
LeBron James
2010
25
CLE
2966
0.506
33.5
41.8
60.4
3.6
Dwyane Wade
2010
28
MIA
2792
0.465
34.9
36.4
56.2
-1.0
LeBron James
2009
24
CLE
3054
0.472
33.8
38.0
59.1
4.1
Dwyane Wade
2009
27
MIA
3048
0.443
36.2
40.3
57.4
-0.5
Tony Parker
2009
26
SAS
2456
0.287
31.7
40.1
55.6
0.2
LeBron James
2008
23
CLE
3027
0.470
33.5
37.3
56.8
-1.5
Dwyane Wade
2008
26
MIA
1954
0.498
33.1
37.5
54.9
-7.0
Tracy McGrady
2007
27
HOU
2539
0.330
35.0
37.5
51.5
-0.5
Dwyane Wade
2007
25
MIA
1931
0.556
34.7
40.5
58.3
-2.2
Allen Iverson
2005
29
PHI
3174
0.432
35.0
37.6
53.2
-2.6
Baron Davis
2004
24
NOH
2686
0.251
30.1
37.9
49.2
-0.6
Seasons with Usage%>30, Assist%>35, Minutes>1200 (1200 was chosen so Rose's season was included)

With the "Rose is not holier than thou" argument out of the way, we can correctly frame him as an all-league player who's top five in a good year and move onto the analysis of the rest of the team. The team's second best player is Joakim Noah, the heart of the team who's an extremely active two-way force. Always dealing with nagging injuries, Noah has averaged 2086 minutes the past four seasons and 64 games, prorating the strike-shortened season to 82 games. Without a quality backup, the Bulls will drop two or three games. Taj Gibson can hold down the fort defensively even at center, but that still leaves a hole in the frontcourt. The other Chicago mainstay is Luol Deng, a Thibodeau favorite for his willingness to work hard and play absurd minutes. Deng showed a couple signs of slipping, and at 28 years old with 22,000 minutes under his belt. His shooting percentages of two-point shots have been low for two straight seasons, and he's been relying more on a subpar outside shot.

Thankfully, the Bulls had a breakout player: Jimmy Butler. The playoffs were a showcase for his aggressive and effective defense, even against a player of LeBron's caliber. His adjusted plus/minus stats were lukewarm, but they heavily rely on priors and don't deal well with players who improve during the season itself. Synergy graded him as a top defender, allowing only 0.76 points per possession against him even though he got a few tough assignments and played in the backcourt next to Nate Robinson. He was tied with Taj Gibson for the best PPP allowed on the team (for players with major minutes.) He should have a full season of elite defense.

The last starter is Carlos Boozer, to the dismay of Chicago fans. He remains a good rebounder with a decent scoring rate, and with so much scoring pressure on him last season his efficiency crashed. Part of that was the inevitable bowing to aging, but his shooting percentages should be boosted by Rose's presence. The best big man off the bench is Taj Gibson, who's somehow already 28. If there was a sixth man of the year for defense, he should have won one. With few offensive skills including a mediocre jumper, he derives his value from tough defense including the team's best shot-blocking rate. If he and Boozer switched minutes totals, the Bulls would be better off, and rumors indicate he'll play heavier minutes. Joining him off the bench is Kirk Hinrich; entering his mid-30's he's not what he once was but he's still a pesky guard at either position with three-point range. The other major addition to the team is Mike Dunleavy, swapped out for Marco Bellineli. He's a skilled, high-volume three-point shooting, averaging plus 5 three's per 36 minutes in recent seasons, giving the Bulls a legitimate outside threat to surround Rose. Marco's stats are similar, but his accuracy has been going south. A tall shooter, he should be effective in Rose-Butler lineups, and his willingness to take a charge will be appreciated with the coaching staff.

With Rose back in full bloom, the Bulls offense will bounce back by +2.5 points per 100 possessions relative to the league average. The defense will also tighten its hold on the opponent, allowing 1.4 less points because, as discussed, Rose is essentially replacing most of the minutes from Nate Robinson and Richard Hamilton and Butler should improve his impact. All told, it's an elite defense joined by a league average offense, good enough for 50 plus wins and homecourt advantage in the first round. But the league has caught onto the Chicago philosophy, and the Pacers will take their place again as the best defense (at least in the east.)

Random prediction: when Boozer goes down with an injury for a few games, the Bulls go on a win streak with Taj Gibson starting.

Win/loss prediction: 52-30


Brooklyn Nets

Six years after Boston sold its future and its youth for Garnett and Ray Allen, the Nets have gambled on the old men of the Celitcs with other new pieces for a shot at a title. At the time of the original big trade, Al Jefferson was a promising young big man and a major trade chip. Years later, he's a big man who no one wants, signing a contract with the NBA equivalent of purgatory. But Garnett is still a defensive juggernaut, holding down Boston's defense as one of the best in the league against all odds. He had one of the largest disparities in defensive efficiency when he was on and off the court -- on the floor, the Celtics defended like the best defense in the league, but when he left they defended like a below average team. While some of this is due to the bad backup effect and the lack of other interior defenders, Garnett has been a plus/minus king for over a decade. In fact, when I worked on the 1997 play-by-play data, he was one of the highest rated players even though he was very young at the time. That's a 17 year-span.

Brooklyn hopes this will be year 18. They've gone deep into their pockets and sold their future first round picks betting on Garnett being able to defend at an elite level another season. Assuaging fears this will be another LA Lakers 2013, the Nets are a deep team with two key players near their primes in Deron and Brook Lopez. Brooklyn essentially kept half of its top ten roster pieces, dumping the aging Gerald Wallace who was a wreck on offense, Kris Humphries and his baggage, backup point guard CJ Watson, young inefficient gunner MarShon Brooks, and Keith Bogans for Paul Pierce, Kevin Garnett, Andrei Kirilenko, Jason Terry, and Shaun Livingston. That five-man swap is one of the best quick rebuilds in recent history.

How will they fit, and how do we know this will be like Boston in 2008, barnstorming the league right away, or LA in 2013, limping to a 7-seed before a quick and merciful out in the playoffs? The Lakers were actually precariously balanced, praying that shot-happy Kobe could co-exist with and defer to the new stars; Nash was supposed to be the one maestro in the world who could operate such a complicated job. But Nash got hurt, and Kobe began to run the team, while Howard recovered from injuries and Pau was a shell of his former self. With a terrible bench the Lakers were doomed. The Nets, by contrast, somehow managed to sign Blatche, a young scoring big man, for a cheap contract, and stunned the league with Kirilenko taking a two-year contract that started at a paltry 3.2 million (he declined a ten million dollar option in Minnesota.) The Russian Mafia Nets are deep enough to rest Pierce and Garnett with minimal damage.

Additionally, this appears to be a team that fits well, depending on Deron Williams. But shots will need to be sacrificed as there's only one ball. In fact, a weighted average by projected minutes of their usage in 2013 with a small adjustment for age found a value of 22.3. Of course, the real average will have to be 20 this season, which means significantly less shots at least for some players. But this isn't necessarily a negative: high usage lineups significantly outperform their projected offensive ratings. For example, using a conservative coefficient, a lineup of with an average usage of 22.3 will translate to +1.8 points per 100 possessions better on offense. Essentially, everyone's efficiency will get a boost from shooting less often.

The counter to that thought experiment is that, yes, on the average level high usage lineups have greater efficiency, but this doesn't work in all cases. But what types of lineups perform well? There are two key factors I believe: passing and spacing. Putting together higher scorers should mean easier shots for everyone, but this doesn't happen without a pass, and spacing will give players room to operate and the spacers themselves create their own opportunities even without handling the ball. Thankfully, the Nets rate well in both categories. The three former all-stars they added are all above average passers for their position, and including the returning stars only the center isn't a superb passer. The team also has plenty of shooting to space the floor as every rotation wing player, with the exception of backup Shaun Livingston, has three-point range, and big men Garnett, Lopez, and Blatche all have midrange jumpers. Kirilenko has a shaky shot for a SF, but he can also masquerade as a PF, and the lone non-shooter is rebounding specialist Reggie Evans who will receive significantly less playing time.

"Rebounding specialist" is an understatement, however. Unexpectedly, Reggie Evans had one of the best rebounding seasons ever. By rebound percentage, Reggie Evans ranked second all-time among qualifying seasons, only being a truly outlier season from Rodman in only 1568 minutes with the Spurs. For years, Rodman led the rebound rate leaderboard, owning the top eight spots. It was a type of domination rarely seen in professional sports. Evans did break one unofficial record though he now has the top defensive rebound rate ever, beating out Rodman in 1995. Only serious NBA die-hards talked about it, but this was simply amazing. Yes, it helped playing next to Brook Lopez, but the Nets were actually an above average team on the boards, thanks in large part due to Evans but also to the rebounding they get from their wings, a true Nets tradition. When Reggie Evans was on the court per NBAWOWY.com, the team rebounded 33.0% of its offensive rebounds and 76.2% of its defensive rebounds (total rate: 54.5.)  Without Evans, the numbers dropped to 28.5% on offense and 71.3% with defensive rebounds (total rate: 49.7.) Basically, with Evans the Nets rebounded at the best rate on both ends of the court, significantly better than any other team, and without him they were a terrible defensive rebounding team and decent enough on offense to keep the overall numbers from being pathetic. It was a Herculean effort on the boards, and when paired with Garnett they should be able to own the glass.


Dennis Rodman
Reggie Evans
Age
Team
Mins.
OReb%
DReb%
TReb%
Team
Mins.
OReb%
DReb%
TReb%
28
DET
2377
16.2
21.7
19.0
PHI
1137
14.1
24.0
19.0
29
DET
2747
15.0
27.8
21.3
TOR
311
11.7
27.7
19.9
30
DET
3301
18.1
34.1
26.2
TOR
798
17.2
34.4
25.6
31
DET
2410
16.1
36.8
26.0
LAC
771
15.3
26.7
20.9
32
SAS
2989
17.9
32.8
25.7
BRK
1967
15.5
38.0
26.7
33
SAS
1568
20.8
37.8
29.7





34
CHI
2088
19.9
33.2
26.6





35
CHI
1947
18.9
32.2
25.6





36
CHI
2856
16.8
31.4
24.1






(It's also important to note these are rebounding estimates: you portion out available rebounds based on how many minutes a guy plays and the average number of opportunities for a rebound a team gets. The numbers in the table above aren't actually the percentage of rebounds grabbed. We can calculate that now for Evans, but since it's unavailable for Rodman I just used b-ref's numbers.)

Beyond Reggie Evans, it was a surprisingly uneventful season for the Nets in their new digs. Deron Williams started the season with awful play, conjuring memories of Shaq showing up to camp fat. But after the all-star break, he was reinvigorated -- his TS% jumped from 54.2 to 62.2, while his usage increased from 23.5 to 26.8. He even had more assists in comparable minutes. Consequently, the team's offensive efficiency jumped by three points, adjusted for strength of schedule. With a fully motivated Deron Williams, the Nets should get another boost people aren't expecting.

One of the most intriguing subplots will be how the large frontcourt of Garnett-Lopez operates. On offense they should do fine, as Garnett is usually right below the three-point line spacing the floor and doesn't demand many low-post touches. But on defense Lopez is slow-footed for a center, much less a power forward, meaning the Nets best defensive player will have to move back to power forward when they share the court. With another year under his belt, there's a possibility he'll be less effective overall if he has to defend the perimeter more. Plus Blatche off the bench might be more suited to certain center match-ups than quick power forwards. However, Garnett's still an intelligent defender with quick feet, and size is always an effective weapon in basketball. Brook Lopez's development, as one of the few young players on the roster, will be important here. He regained his rebounding mojo, and on the offensive boards he's actually pretty good. Garnett offsets his weaknesses in defensive rebounding and pick-and-roll defense, and the Nets may find a lot of value in a Garnett-Kirilenko frontcourt, but Brook Lopez is the best scorer for centers in the league. He has a decent jump shot, but performs well in pick-and-rolls and has a low post game with a decent touch. Unlike other behemoths, sending him to the line isn't an option, where he's at nearly 80% for his career. Few modern centers can score like he can, and given his age he may take another step forward under the tutelage of Garnett.

With six former all-stars, the Brooklyn Nets project as a dangerous team with multiple weapons. Their age will result in heavy rest for their oldest stars, but their depth is strong enough to weather the regular season. They could be an even more dangerous playoff team, able to play Garnett and Pierce major minutes, and could attack teams with big lineups (Joe Johnson-Pierce-Garnett-Lopez) or effective smallball lineups with Kirilenko-Garnett. The Nets last season lacked defense and offensive support on the wings. They've added one of the greatest defensive players in history and replaced Gerald Wallace with Pierce and Kirilenko. Most analysts are unimpressed with their haul for this season, predicting only 50 to 52 wins, but they should win more games just from their new SF rotation, much less Garnett, and they could be even more dangerous in the playoffs.

Random prediction: Kirilenko narrowly misses sixth man of the year with many writers upset he was ignored on some ballots in favor of one-dimensional scorers.

Win/loss prediction: 56-26


Indiana Pacers

Since few people paid attention to the Pacers at the beginning of the season, few saw the remarkable recovery and evolution from Hibbert. After the criticism of signing a large extension with Hibbert in the offseason, Indiana watched as the big man clanked brick after brick. The media penned it as the inevitable fall once a young player gets big money, and that Hibbert was a stiff who never deserved the money in the first place. Well, at some point in the season, people noticed the Pacers were defending at an elite level, and Hibbert was the magnificent backline anchor. Then Indiana's offense was slowly getting better, and by the time they met the Heat in the playoffs they were at full strength, Roy using his immense size to wall off the paint while feisty defenders George and Stephenson harassed LeBron. They took the champions to an improbable seven games before bowing out. What happened?

Hibbert suffered a wrist injury before the season, one that was under-reported save for a few writers. He was shooting 41% before the all-star break, woeful enough for a small perimeter play who relies on outside shots, much less a 7' 2" big man -- he was shooting 43% near the basket, per Mahoney's numbers. Yet all his other numbers were virtually the same as his historical averages -- his rebounds, his turnovers, his steals, and his blocks had actually improved. His usage was down (in the first half of the season) a little, but his efficiency had plummeted to horrendous levels. However, once he recovered, the Pacers offense bounced back from the doldrums to rates above the league average. Given their stifling defense, a renewed Hibbert led to a powerful team, and one that was underestimated in the post-season. Before the all-star break, their SRS (point differential with a simple strength of schedule adjustment) was only 2.4. After the break? 5.7, which is basically the difference between a 56 win team and a 47 win team. And the improvement was all offensive, as their defensive efficiency actually slipped a small amount. Before the break, their offensive efficiency was 102.6; and after it was 107.0.

If you track Hibbert's progress with the team's offensive efficiency, there are clear correlations. The first graph below shows Indiana's offensive efficiency in green progressing throughout the season. A ten game average is used to show the gradual change in their offense. Hibbert's scoring, in terms of points scored per 75 team possessions (this is roughly 40 minutes at Indiana's pace, roughly speaking), is shown in blue. It steadily climbs with the team until the end of the regular season, dips in the middle of the playoffs, and then skyrockets in the series versus Miami. His efficiency tracks just as well. The second graph is showing PTS/(FGA+.44*FTA+TOV)*100. It's like TS% but with turnovers included. Again, he numbers climb as the season progresses, and he peaks again versus Miami.




Any projections that base the strength of the Pacers on their entire season are missing this important trend. The Pacers were actually the 8th best offensive team, via b-ref's ratings, in 2012, and while a lot of that is due to Granger they shouldn't slip all the way to 20th on offense the next season. Expect the Pacers to hang around the league average offensive efficiency mark, and if Granger is healthy they may eclipse it by a decent amount.

Besides the big man in the middle, the other story of the season was how terrible their bench was. When their starting lineup was in the game they outscored opponents by 12.1 points per 100 possessions. Otherwise they only outscored opponents by 1.7 points. Their lead bench players were Tyler Hansbrough, Gerald Green in a disappointing season, the limited Ian Mahinmi, and the short and defensively combustible DJ Augustin. Bringing in Scola, CJ Watson, Copeland, and hopefully more minutes from Granger, they should get a big boost in depth. One of the problems with judging a roster by your eye is that you miss the differences between the minor players. Few people care about CJ Watson, but he's a much better option than Augustin, who couldn't defend even with all this support.

Their bench segues into the biggest question of 2014 for the Pacers: what will they get out of Danny Granger? A former all-star and a high scorer, even if he can't get to the rim reliably he's a large volume three-point bomber with good accuracy. He's a surprisingly good defender too, though with all the injuries he may have problems with quick players. He's at least a better option than Gerald Green though. Unfortunately, he's already out three weeks with an injury, and Indiana may decide to use him as a trade piece to net someone more serviceable. In any case, there's a good chance he provides the team with some value

Even with the improvements in the east, the Pacers have a young core, upgraded their bench, and should be better on offense than last season due to Hibbert's recovery. It's enough to move into the mid- to high-50's win range and set themselves for a healthy seed come playoff time, where we won't underestimate them again.

Random prediction: Paul George makes second team all-NBA, over players like Carmelo Anthony.

Win/loss prediction: 56-26


Miami Heat

Well, we all know how that season went, right?

After a post-title nap, the Heat awakened in the middle of the season and tore off 27 wins in a row. They hit some speed bumps in the playoffs, however, and squeaked by the Pacers and Spurs. But what really happened? And are they the 27 win team or the one nearly defeated by Indiana? During the historic win streak, the Heat were remarkably lucky in close games, and they ended the season with 66 games even though by point differential their strength suggested a 62 win team. The win streak was vastly overrated as well; you don't get bonus points for winning games in a row. They still count as one each, and it just means you're losing during a different part of the season. That's indicative of an inconsistent team, which is worrisome. And for good reason -- the Heat acquired Chris Anderson before ripping off the win streak, but the other main component was Wade finally playing healthy. But he was up and down during the season, and had the worst playoffs of his career. He shot under 50 TS%, his shot volume sank, and teams started to ignore him on the perimeter. The Heat were crippled, and LeBron barely dragged them to their second title. People contend since the Heat barely won the title, they aren't strong champions and won't win three in a row. But they fail to mention the most important factor that no one controls: Wade's health. He's reportedly in great shape, but we won't know until a few weeks into the season to see how his body holds up.

Picking up Chris Anderson after he suffered a bizarre problem with the law, the Heat masked a weakness with the center's excellent rebounding and weakside defense. He set an NBA record for field-goal percentage in the playoffs, operating out of his "bird cage" by standing a few feet from the basket and either crashing the boards or cutting to the hoop for an easy basket once his defender moved toward LeBron. The Heat were a poor rebounding team, even with Anderson for half the season, ranking 27th and 24th, respectively, in offensive and defensive rebounding percentage. While the space-and-pace strategy necessitated a perimeter player like Battier as their power forward, they conceded size inside and rebounding, which nearly led to their downfall versus the Pacers.

Anderson could only help with rebounding and some aspects of defense, however, as he was outmuscled in the paint by Hibbert. But their new addition Greg Oden was tailor-made to address their weaknesses. If he can make it through the season unscathed, which he has yet to do, the Heat have a major counter to use in the playoffs versus big teams. Even though he looked like a newborn fawn in the pre-season and may have lost a lot of quickness, his immense size and strength will give the Heat what they've been missing. The center position for the Heat has been the Spinal Tap's drummer of NBA positions. They've cycled through numerous players trying to find the right player, usually just relying on natural power forward Bosh in the middle and giving up defensive rebounds. As respected as the front office is (though it doesn't take a genius to think Wade+LeBron+Bosh is a good team), they've whiffed on their other role players before with a strange fascination with players like Eddy Curry. If Curry had been healthy and played at his best, it would mean the Heat would gain a center who (again, at this best) was a terrible defender, a historically bad rebounder, and an inefficient scorer who clogs the paint. With whatever the Heat needed, how could Curry provide any of those things? Oden is the exact opposite, a gamble with a high risk but a high payoff and addresses their weaknesses. But Beasley represents the front office that thought Curry was a good idea. If the Heat are lucky, Beasley receives no playing time, as he's a moody, inefficient gunner who offers nothing else of value besides inaccurate long two-pointers. He offers zero to Miami's philosophy: defense, spacing, and transition attacks, much less shoring up their problems with rebounding and interior defense.

The reason why you don't want someone like Beasley taking shots away from LeBron is that LeBron had an absurd shooting season. With great shooting percentages on his jumper and his typical dominance in the paint, he somehow managed to raise his FG% for the sixth straight season. The only other player in history who shot 56% from the floor with over 100 attempts from outside was Charles Barkley, who shouldn't have been shooting them anyway. LeBron hit the "efficient frontier" of high volume seasons, the outer most edge of efficiency for how much he shot. He scored 116 more total points than Carmelo even though the Knicks forward took 135 more field goals -- and it was one of Carmelo's most efficient seasons. The Heat's offense was elite, one of the best post-merger. The kicker? LeBron's was only the second most efficient season in the league, a hair behind Durant. But the Miami Heat led the league in TS% over the Thunder, not just thanks to LeBron but Wade and Bosh, even though Bosh took over 500 shots past 16 feet.


The Heat would have ranked even higher, due to their efficient role players like Ray Allen, Mike Miller, and Chris Anderson, if it weren't for Norris Cole, who had a TS% of 48.3 off the bench. He also had one of the worst adjusted plus/minuses in the league, even though Heat fans began to love his energy and hustle. The Heat nearly topped the unofficial record for team TS%, which stands at 58.9 from the '85 Lakers. If Cole's shots had been replaced by someone shooting the average mark, Miami would have topped 59. Beasley and Cole exist to give the opposition a rest from the relentless attacks of LeBron, Wade, Bosh, and the three-point barrage behind them. Nevertheless, Cole's pesky defense is useful off the bench, and if his efficiency improves he'll be a fine reserve guard.

The rest of the bench is led by Ray Allen, who will be nearly the oldest player in the league but will be relied upon for his outside shooting and (sometimes) ballhandling. These days half of his shots are three-pointers, but the Heat don't mind. His defense is fairly good for a player his age, but Miami keeps him away from anyone who can do harm. Shane Battier is the other major rotation piece. People wrongly labeled LeBron a power forward because the Heat played smallball, but Battier was the real power forward. He usually guarded the bigger forward and spent more time on defense in the post versus guys like David West. He was essentially a stretch 4, waiting in the corner for a pass. Battier's a larger guy than people realize too at a legitimate six eight. People may argue that LeBron's size and rebounding make him the power forward, but if he didn't guard power forwards on defense, led his team in assists, and slashed to the rim and hit three's, why would he be a frontcourt player?

A Miami Heat player his whole career, Haslem will be called into starting duties for another season. He's turned into the coach's veteran favorite who probably receives too many minutes for his dissipating talent. He took over the job in the middle over Joel Anthony, a Canadian center with exceptional and quick defense but no ability to rebound and one of the worst offensive games in the league. The other starter, besides one of the big three, is Mario Chalmers, who doesn't run an offense well but doesn't have to, as his gifts as an outside shooter and ballhawk are a perfect fit for the team. And believe it or not, but he's older than Greg Oden. James Jones will receive more playing time without Mike Miller, but entering his mid-30's he could be exposed and the Heat will have to pray Battier stays healthy.

The Heat saw little roster change, only the additions of Oden and Beasley while Mike Miller got the amnesty ax, but all three players will likely only get small roles. With few changes and the dependability of LeBron and Bosh, the season hinges on Dwyane Wade. When he's healthy, the Heat are a scary team and stronger than anyone in the league. When he's not, he's an aging shooting guard who can't shoot, and you don't' want to take the ball out of LeBron's hand, so teams start to ignore Wade on the perimeter. The Heat are still good enough even without a healthy Wade to win a large number of games, as they showed in the playoffs, but if they want a third title in a row, in the fourth finals appearance for this group, Wade has to be healthy when it matters. If the Heat were smart, they'd treat Wade like the Spurs do Ginobili and Duncan, keeping him fresh for the end of the season. LeBron and Bosh are more than enough offensive firepower on the court, and the Heat only need to hit their sixth gear versus the strongest of teams.

After a season with 66 wins, 27 of which were in a row, and a second title, they're likely to regress. Miami was lucky to win 66 games, and they should see a major drop this season. But if you use their expected wins from point differential, 62, and factor in the improvements in the east, it's only a small slip that's mostly due to age. If Wade is healthy for the playoffs, they will have a good chance at winning their third title in a row, which is actually a rare event in the NBA -- Jordan spoiled us, as it's only been done by Jordan's Bulls, Shaq's Lakers, and Russell's Celtics. LeBron will have plenty of motivation for this title because he would separate himself historically from all but two legends. With a bevy of contenders and several elite teams with major changes or question marks, it's a year for the field with crushing odds to take the title -- but as a single team, the Heat are still the most likely.

Random prediction: Oden stumbles through the regular season, the team being overcautious and taking him out for a week at a time due to the slightly breeze, but he finally sees playoffs minutes.

Win/loss prediction: 58-24